Our pipelines sometimes get stuck (example).Some workers get decommissioned due to spot termination and then the new workers get added. However, after (1) Spark doesn't notice new executors: And I don't know why. I don't understand how to debug this,...